Effect of Selection for Age at Time of Puberty On ...

r

STATION TECHNICAL BULLETII 25 JULY 1952

Effect of Selection forAge at Time of Puberty

On ReproductivePerformance in the Rat

EDWARD P. WARREN

RALPH BOGART

viaRAOCT 3 1952

Agricultural Experiment StationOregon State College

Corvallis

ACKNOWLEDGMENTS : Statistical aspects of this study were conducted incooperation with the Bureau of Animal Industry, United States Department-of Agriculture, and the State Experiment Stations under the Western RegionalProject on Beef Cattle Breeding Research. Thanks are due Drs. J. C. R. Li,Hugo Krueger, and Paul Bernier for statistical counsel-

Effect of Selection for Age at Time ofPuberty on Reproductive Performance

In the RatEDWARD P. WARREN and RALPH BOGART

Department of Animal Husbandry

D EVELOPING AND MAINTAINING maximum reproduc-tive efficiency in farm animals is important to the economy of

the livestock industry. Physiologists, geneticists, nutritionists andothers have worked toward this goal in recent years. The problem iscomplex, involving the combined effects of numerous genetic andenvironmental factors. The solution can be obtained only throughdetailed study of each of the forces involved.

Except for poultry relatively little has been done on sexualdevelopment, particularly with reference to the over-all effect ageof pubescence may have on fertility. Neglect of this approach to theproblem probably stems from the belief that pregnancy and subse-quent lactation early in the life of the individual resulted in a reduc-tion in mature size, fertility, and quality of offspring. It has beenrecently shown (McKenzie, 1928; Bogart et al, 1940), however,that early-bred and lactating swine and rats reached the same adultsize as those bred at normal age. It was noted that early and regularreproduction stimulated continued reproduction. One explanation(Asdell, et al, 1941) is that, reproduction results in greater unitybetween gonads and other endocrine glands which regulate repro-ductive processes. This line of reasoning leads to the hypothesisthat some factors affecting rate of sexual development may be as-sociated with reproductive performance. Asdell (1946) has statedthat such is the case in several species and that early puberty andhigh initial fecundity may be good bases for economic selection.

If age of pubescence is used as a practical criterion for selec-tion, information must be available on effectiveness of selection forthat character (heritability) and genetic limits of variability. Know-ledge of the physiology of sexual development also would be valu-able. This need made it advisable to conduct a detailed study ofgenetic and endocrine factors affecting sexual development and therelationship of rate of sexual development (age at time of pubes-cence) to some productive characters in the female.

The rat was used as the experimental animal since large numberscan be reared in a comparatively short time with minimum expense.Genetic studies utilizing rats as experimental animals permit experi-mental designs more adaptable to statistical analyses.,43

STATION TECHNICAL BULLETIN 25

Review of LiteraturePuberty

Puberty is the time when reproduction is first possible, i.e.,when germ cells are released, (Asdell, 1946). Vaginal opening,changes in the vaginal smear, sexual receptivity, and ovulation areintegral factors in attainment of puberty by the female rat (Blandauand Money, 1943). Time of establishment of these factors is notnecessarily identical. Long and Evans (1920) found that in themajority of their females, vaginal introitus preceded ovulation byabout 5 days. Blandau and Money (1943) reported that vaginalopening generally occurred between 48 hours before and 12 hoursafter the onset of heat. These same investigators observed first ovula-tion in the majority of females between 5 and 10 hours after theonset of first heat. They also reported that 30 per cent of the femalesmated during the first heat period and 19 per cent of those insem-inated at first heat failed to conceive. A small percentage of femalesdid not become sexually receptive even though vaginal opening andovulation had occurred. These data indicate that no single phenome-non is a wholly reliable criterion of time of puberty. Since all ofthese phenomena usually occur within one day, (Asdell, 1946) noserious error is involved when one relies on any of these.

The age at time of puberty in rats is quite varied. Long andEvans (1920) found that their strain of rats reached puberty at anaverage age of 76.5 days. The range was from 53 to 142 days.Freudenberger (1932) stated that the Wistar strain attained pubertyat 46.9 days after birth, with a range of 36 to 66 days. In his labora-tory the Long-Evans rats reached puberty at 52.7 days with a rangeof 39 to 101 days. Slonaker (1924) gave the average age at time ofpubescence in his strain as 80.6 days. The average age at time ofvaginal opening of different groups of intact normal rats at Cornellwas observed to be between 58.6 and 61.1 days (Bogart, et al, 1940).The reports of other investigators (Greenman and Duhring, 1923;King and Donaldson, 1929; Engle, et al, 1937; Blunn, 1939; andStone and Barker, 1940) give further evidence of widespread differ-ences in age at time of puberty.

Factors affecting sexual development

An unknown, but relatively small, portion of this variation inage at time of pubescence is undoubtedly due to the different criteriaused by the various investigators in measuring time of puberty.Different environmental conditions existing in the several laboratorieswhere studies were carried on is another source of variation. The

REPRODUCTIVE PERFORMANCE IN THE RAT

fact that a considerable part of the variability is due to genetic causesis borne out by the different values obtained for different strains inthe same laboratory (Freudenberger, 1932 and Blunn, 1939). Byselection within a strain Stone and Barker (1940) were able toestablish two divergent lines with regard to age at time of pubescence.

Another point of interest is the observation that the reportedage at time of puberty has shifted downward by as much as 30 daysover the last 50-year period. Stone and Barker (1940) attributed thischange to improved environment (diet and management), to use ofdifferent criteria of measurement and to unrecognized causes, one ofwhich may be selective breeding in which the genetically late pubes-cent animals have been unknowingly eliminated.

AGE AND WEIGHT: The work of Asdell and Crowell (1935)and studies of Bogart, et al, (1940), involving observations on 300rats indicate that the time of vaginal opening is more closely relatedto age than weight. The coefficient of variability for age at vaginalopening, as reported by the latter group of workers, was 4.89 percent while that for weight was 10.55 per cent. The coefficient of cor-relation for age and weight at the time of vaginal opening (afterremoval of variance due to treatment, location of animals and inter-action) was 0.57. Blunn (1939) also found that time of vaginalopening was influenced by age to a considerable extent. Cole andCasady (1947) reported that their earlier maturing strain of ratswas heavier than the late maturing strain at 85 to 90 days of age.

RATE OF GROWTH: The work of Bogart, et al, (1940) indi-cates a marked correlation of growth rate with time of vaginal opening.These workers found that time of vaginal opening in a group of 50spayed females occurred at the same point in the growth curve as itdid in intact rats. These authors interpret this as an indication thatcertain of the factors affecting sexual development also governgrowth rate. Asdell and Crowell (1935) studied effect of retardedgrowth on age at time of vaginal opening in the rat. They subjected40-gram rats to severe low calorie inanition. Ages at time of vaginalopening in three such rats were 377, 101, and 106 days. Theirweights at the time of vaginal opening were 42, 43, and 54 grams,respectively. Average age at time of vaginal opening for the controlgroup was 53.5 days. The interpretation was that growth and sexualdevelopment were correlated by the hypophysis, a secretion of hor-mones stimulating sexual development, which increased as the pro-duction of growth stimulating hormone decreased. Evans and Bishop(1922) carried out similar experiments with like results. Blunn(1939) computed average growth rate from weaning to the time


of vaginal opening. He found a significant correlation betweenaverage growth rate and age at time of vaginal opening and con-cluded that vaginal opening was related to growth rate and age.

SIZE OF LITTER AT WEANING: Engle, Crafts and Zeithaml(1937) found a definite relationship between size of litter raised toweaning and age at time of vaginal opening of females in the litter.Females reared in smaller litters reached puberty earlier than thosein large litters. If these findings are correct, selection for earlypuberty should result in smaller litters. Blunn (1939), however, wasunable to detect any significant effect of litter size at weaning on ageat time of vaginal opening of females in the litter.

NUMBER OF YOUNG BORN : Little information is availableon number of young born. If number raised to weaning is significantin determining time of puberty, number of young born would beexpected to exert a similar, but lesser, effect since the former is adirect function of the latter. This factor might affect secondary cor-relations, such as correlation of number born with suckling ability,which would be reflected in growth rate. Such an effect probablywould not be of consequence. A direct genetic correlation betweennumber born and time of vaginal opening is theoretically possible.

No studies with rats or other mammals have been reportedin which an attempt was made to determine heritability of age ofpubescence. It is possible that such an estimate might be calculatedfrom the original records of Blandau and Money (1943).

Relation of sexual development to productive charactersLittle evidence is presented in the literature relative to the

effects of selection for early puberty on growth, mature size, fer-tility, length of reproductive life, suckling ability and other pro-ductive characters. Some definite evidence of a correlation betweenage at time of puberty and growth has been previously cited (Bogartet al, 1940). Cole and Casady (1947) also found that their earlier"maturing" strain of rats was significantly heavier, at 85 to 90 daysof age, than their later "maturing" strain.

Bogart et al, (1940) found no evidence of a relationship be-tween age at time of puberty and mature weight in rats. Phillipsand Zeller (1943) on the other hand found that age at time ofpuberty was significantly different in the large and small types ofPoland China swine.

One might logically postulate a correlation between sexual de-velopment and fertility (litter size). Both are subject to the sameendocrine regulation even though sexual development is qualitative


and fertility is quantitative. In this connection, Cole and Casady(1947) found highly significant differences in both age at time ofpuberty and prolificacy between two highly inbred strains of ratsderived from the same parental stock. Of perhaps equal importancewas the significantly greater ovarian weight of the early "maturing"more prolific strain. Hetzer and Brier (1940) found that the larger,earlier "maturing" type of Poland China swine was also more pro-lific than the small, late "maturing" type. MacArthur (1944) select-ed mice for large and small size for several generations, and althoughhe did not determine the effects of such selection on age at timeof puberty, he did find that the larger strain produced significantlymore young per litter. This was determined to be due primarily to alarger number of ova shed. He concluded that the correlation be-tween body size and litter size depends on endocrinal links, such asthe growth and gonadotropic hormones of the anterior pituitaryrather than on differences in "litter size genes" or "fertility factors"as such in the two strains. While one may not hold exclusively toMacArthur's explanation, endocrine factors are unquestionably ofconsiderable importance in such relationships. Accordingly, one canreadily conceive of similar "links" between such characters as ageat time of puberty, length of reproductive life and suckling ability.

Material and MethodsSelection studies

Forty-two female and nine male albino rats of the Sprague-Dawley strain were obtained from the Holzman Rat Company,Madison, Wisconsin, on March 21, 1950. The females, when re-ceived, were 43 days of age and all prepuberal, whereas the maleswere approximately three and one-half months of age and readyfor service as breeding animals. The data presented are the recordsof performance of these individuals and three generations of theirselected descendants reared between March 21, 1950 and May 1, 1951.

Selection for early and late pubescence was used to establishtwo divergent lines. They will be called the "early" line and "late"line. Ten females and four males were selected from first littersin each line as parents of the next generation. Selection of breedingfemales was based solely on age at time of pubescence determinedby date of vaginal opening. Selection of breeding males was basedon age of vaginal opening of female littermates. One individual wasselected at random from each of four litters. In the foundationstock, selection of males was entirely at random. Females were al-lowed to reproduce at regular intervals until death, sterility, or termi-


nation of the experiment interrupted. No attempt was made to re-place breeding animals that died or were infertile, regardless of pos-sible cause.

Every effort was made to maintain uniform environmentthroughout the experiment. Experimental animals were housed inan unlighted, well insulated room maintained at a temperature of24 to 27 degrees centigrade. Heat was supplied by an electricallyoperated, thermostatically controlled, forced air heater.

Dornfeld's rat mix (Table 1) and water were fed ad libitum.Feed cups and water bottles were checked daily. Fresh pork liverwas provided a the rate of 12 grams weekly for mature, nonlactatingfemales and mature males. Lactating females received a minimumof 15 grams and immature rats received proportionate amounts perweek depending on age.

Table 1. FORMULA FOR DORNFELD'S RAT MIX.

IngredientUnits ofweight

Yellow corn meal (whole corn) -- 38.0Whole wheat meal ....-------- 32 0Powdered skimmed milk ----------------------------------------------------------------------------- 20.0Whole ground alfalfa meal --------- 60Cod liver oil ------------------------------------ 2.0Yeast (Fleischman's dry type) 1.0Calcium carbonate or lime ----- 0.5Sodium chloride ---------------------------- 0.5

Total 100 0

Mating was accomplished by placing the female in the cagewith the male for 5 days. She was then removed for 3 days. Thisprocedure was repeated, using a different male at each exposureperiod, until it was evident by palpation that the female had con-ceived. With this method, it was possible to identify the sire of eachlitter without determining time of estrus in each female by dailyvaginal smears. The method has the objection that some estrualperiods occurred during the nonexposure period of 3 days and con-ception was delayed a few days beyond the earliest possible date.This is of no particular consequence in the present study. All fe-males were first exposed to the male at 100 days of age and at 31days after delivery of each litter, regardless of whether the littersurvived. This procedure provided for a 6-day rest period aftereach litter was weaned. No attempt was made to mate particular

9

I


females to certain males, except that brother-.sister matings were notmade for production of first litters.

Records taken were date of birth of each litter ; number born inlitter ; at 25 days of age, the number weaned and sex ; and individualweaning weights. After weaning, daily observations were made onthe females for vaginal opening. Date and weight at the time ofvaginal opening were recorded for each rat. A few females werenot observed for vaginal opening as they were directed to otherexperiments. Weights were taken on all first-litter females at 75days of age.

Statistical analysesDifferences between the two divergent lines in selection of

studies were analyzed by analysis of variance (Snedecor, 1946).Relation of puberty to productive characteristics was determinedby multiple regression after variance and covariance due to line andto litter sequence were removed (Snedecor, 1946). Heritabilityestimates for age at time of vaginal opening were determined direct-ly from the line differences (Krider, et al, 1946).

ResultsHeritability of age at time of vaginal opening

The mean age at time of vaginal opening for first litter femaleprogeny of each line in each generation is noted in Table 2. Dif-ference between means of the two lines for each generation, also inTable 2, was 0.58, 0.95 and 2.25 days for the first, second, andthird generations. These are estimates of genetic difference accu-mulated as a direct result of selection in preceding generations foropposite extremes of age at vaginal opening.

Table 2. MEAN AGES (IN DAYS) AT TIME OF VAGINAL OPENINGFOR ALL FIRST-LITTER FEMALE PROGENY BY LINES

AND GENERATIONS.

Generation

Foundation First Second Third

LineNum-ber

Mean Nurn- Meanage ber age

Days Days

Num-her

Meanage

Days

Num-her

MeanageDays

Late ------------------ 143 43.74 27 44.08 32 41.88

Early ................ 435

55.14 38 43.16 23 4313 30 39.63

Late minusearly ...... 0 58 0.95 2 25

- MEAN OF ALL FEMALE PROGENY--- MEAN OF DAMS OF NEXT GENERATION

54

524

50

48

46.

44

42.

40

381T -r30

3.48;

(225);

10 STATION TECHNICAL BULLETIN 25

LATE LINE --( above)

zzW0-0 2.93

EARLY LINE(below)

2GENERATION

15.74 13.07 21.81

1-ERITABILITY (per cent)

Selection for early and late vaginal opening in each generationand line is represented by average superiority or inferiority of thedams compared with the means of the population from which theywere selected. The total amount of selection for each generation isrepresented by the sum of the amount in the two lines. Thesevalues are presented in Table 3.

In the first generation the mean inferiority of the late-line damof that generation was 4.44 days and the mean superiority of theearly-line dams was 2.93 days. Average age at the time of vaginalopening of the selected late-line dams was 59.58 days (Table 4).Average age at the time of vaginal opening of the population fromwhich they were selected was 55.14 days (Table 2) or a differenceof -4.44 days. Similarly, the average age at the time of vaginal

P

REPRODUCTIVE PERFORMANCE IN THE RAT 11

Table 3. MEAN SUPERIORITY (+) OR INFERIORITY (-) OFSELECTED DAMS IN AGE AT VAGINAL OPENING IN DAYS.*

First

Dams of generation

Second Third

Line Number Mean Number Mean Number Mean

Days Days Days

Late ------------ 10 -4.44 8 -3.41 7 -348Early _......... 9 +293 6 +3.77 7 +2.60

Late minusearly ........ 7.37 7.18 6.08

* In calculating the mean of the dams of each generation, each dam was weightedaccording to its number of female progeny. The mean for the whole population from whichdams were selected was then subtracted from the mean for selected dams to obtain theaverage superiority or inferiority of dams.

opening of the selected early-line dams was 52.21 days (Table 4)while that of the population from which they were selected was 55.14days (Table 2) or a difference of +2.93 days. Total selection in thefirst generation thus amounted to 7.37 days (4.44+2.93). Corres-ponding values for the second and third generations were calculatedin the same manner. In calculating the mean age at the time ofvaginal opening of the selected dams in each line of each generationit is necessary to adjust the value for each dam according to thenumber of female progeny raised. The adjusted means of the se-lected dams of each line and generation are presented in Table 4.

Table 4. MEAN AGES, IN DAYS, AT TIME OF VAGINAL OPENING OFSELECTED DAMS BY LINE AND GENERATION.

Dams of generation

First Second

Mean Mean

Third

Meanage

Days47.5640.53

Line Number age Number age Number

Days Days

Late 10 59.58 8 47.15 7

Early 9 52.21 6 39.39 7

Estimates of heritability (h2) are based on the association ofthe selected dams with their offspring. Estimates of heritability ofage at the time of vaginal opening in the present study were calcu-lated from the values given in Tables 2 and 3. Thus h2 for a givengeneration equals the difference between the means of the progeny

C

4

P

4

Heritability estimates for age at the time of vaginal opening were16, 13, and 22 per cent for the first, second, and third generations,respectively, (Table 5). The average heritability for the three gen-erations was 17 per cent (Table 5). A diagram showing the amountof selection in each line for each generation and the accumulated

0

f


of the two lines for that generation divided by the total accumulatedselection. It is necessary to multiply this value by two since it isassumed that the males used were representative of the breeder popu-lation used in each generation which would then mean an equalselection differential on the male and the female side. Thus for thesecond generation :

0.95 (Table2) X2h2 = =0.1307 or 13.07 per cent

7.37 + 7.18 (Table 3)

genetic difference between lines during the three generations is pre-sented in the chart on page 10.

Table 5. SUMMARY OF HERITABILITY ESTIMATES OF AGE AT TIMEOF VAGINAL OPENING.

eneration

Differ-ence

betweenlines

Accumu-lated

selection h2 2

First ......................................................Days0 58

Days7.37 .08 .16

Second ------------------------------------ 0 95 14.55 .07 .13Third ............ 2.25 20.63 .11 .22

Average ----------------------------------------- .17

Differences between early and late linesSelection for extremes in age at time of vaginal opening re-

sulted in age differences between early and late lines for time ofvaginal opening and certain productive characteristics.

AGE AT TIME OF VAGINAL OPENING: The mean age atwhich vaginal opening occurred was 41.9 days for early-line progenyand 44.4 days for late-line progeny, a difference of 2.5 days earlierfor the early line (Table 6).

The difference in average age at which vaginal opening occurredin females of the two lines was significantly different from zero(Table 7).

Tine

7'

Degreesof

freedomM canszquare F

1 629.53388 11.89

9"I

III


Table 6. AVERAGE AGE AT TIME OF VAGINAL OPENING.

Average age atNumber female vaginal opening

DaysEarly 190 41.9Late 200 444

Difference ........................................

Table 7. ANALYSIS OF VARIANCE FOR AGE AT TIME OFVAGINAL OPENING.

Source of variation

Between lines ---------------------------------------------------------- 52 96*Within lines ------------------------------------------------------------

Total -------------------------------------------------------------------

" Significant.

389

01, NUMBER OF YOUNG BORN PER LITTER : The average numberof young per litter born to females was 9.66 for the early-line and8.69 for the late-line (Table 8). These means are based on com-bined production of foundation females and first and second genera-tion females of each line. Mean litter size for early and late-linefemales for each generation (0, 1, and 2) and the difference betweenmeans is given in Table 8. Average litter size was 9.27 for early-linefoundation females, making 0.74 more young born per litter in theearly-line. For first generation females the mean litter size was9.84 for the early-line and 8.69 for the late-line; a difference of 1.15

Table 8. AVERAGE NUMBER OF YOUNG BORN PER LITTER.

Generation

Foundation First Second Total

Num- Num- Num- Num-ber Young her Young ber Young ber Youngof per of per of per of per

Line litters litter litters litter litters litter litters litter

Early ---------------- 44 9.27 25 9.84 10 10.90 79 9.66Late .................. 45 8.53 29 8.69 10 9.40 84 8.69

Difference ---- 0.74 1.15 1.50 0.97

If

M1

Si

1116_

LL

III

ti

If atmo a I'I iup--;,R u

V9 9 191

aznbs upaylt

inopas; 30 s o

M1

F

1

£I'9£


young per litter. For second generation females the difference be-tween lines was 1.50, the mean litter sizes being 10.90 for the earlyline as compared to 9.40 for the late line.

Variation between mean litter sizes of the early and late lineswas significantly different from zero when the production of thethree generations, involving a total of 163 litters, was combined(Table 9).

Table 9. ANALYSIS OF VARIANCE FOR NUMBER OF YOUNG BORNPER LITTER.

Between lineWithin lines

Total .......

* Significant.

Source of variation

b, NUMBER OF YOUNG WEANED PER LITTER : The average num-ber of young weaned per litter, omitting litters in which none of theyoung survived, was 8.26 for the early line and 8.13 for the late line(Table 10). These means are based on production of the foundation(0 generation) females and first and second generation of the twolines.

Table 10. AVERAGE NUMBER OF YOUNG WEANED PER LITTER ANDPER CENT OF ALL LITTERS IN WHICH No YOUNG SURVIVED.

Numberof Not

litters weaned I litters raised

Per cent

Early 53 27.4Late 57 26.9

Difference ............... 0 07 05

The percentage of total litters in which none of the young wereraised was similar for the two lines. Early-line females failed tosuckle 27.4 per cent of their litters and late-line females failed tosuckle 26.9 per cent (Table 10).

Between lines ...................... ----- ------ ---------------Within lines ....--

Total . _ .............------.--------------.-..-. ......

Line

1

108

10'9

0.13

Numberof

females

F

0.02

4.4

Degree,of

freedom

1

,388

3 899

Meansquare

1,&48.84SSI)4

34..3'

6.93


The difference between lines with regard to the mean numberof offspring per litter raised to weaning (25 days) was not sig-nificantly different from zero (Table 11).

Table 11. ANALYSIS OF VARIANCE FOR NUMBER OF YOUNGWEANED PER LITTER.

Source of variation

Degreesof Mean

freedom square

AVERAGE WEANING WEIGHT: The average weaning weight(25 days) of the 190 early-line female offspring was 46.9 gramswhile that of 200 late-line female offspring was 51.3 grams, thus thelate-line offspring were 4.4 grams heavier (Table 12).

Table 12. AVERAGE WEANING WEIGHTS

EarlyLate ------------

Difference

WeightGrams

46.951.3

The average weaning weights of the female offspring of thetwo lines were significantly different (Table 13).

Table 13. ANALYSIS OF VARIANCE OF AVERAGE WEANING WEIGHTS.

Between linesWithin lines

Source of variation

Total -----------------------------------------------.--------------.------

* Significant

Line

Vi

200

Weight

Gaou11127126.09

14.82

r

Line

freedom square

389

Numberfemales

F

98.36-lines ...lines

...............................................................

3.9


WEIGHT AT TIME OF VAGINAL OPENING: Late-line femaleprogeny (first, second and third generations) averaged 14.82 gramsheavier at the time of vaginal opening than early-line female progeny.Average weight at the time of vaginal opening was 126.09 grams forlate-line progeny and 111.27 grams for early-line progeny (Table 14).

Table 14. AVERAGE WEIGHT AT TIME OF VAGINAL OPENING.

Numberof

females

Early ............. ...... ....................................... -------- --Late ..............................................................................................

Difference ------------------------------------------------------------------------------

Average weight of female progeny at the time of vaginalopening was significantly different between the two lines (Table 15).

Table 15. ANALYSIS OF VARIANCE OF WEIGHT AT TIME OFVAGINAL OPENING.

Degreesof Mean

Source of variation

Between 21,389.44Within 217.46

Total

' Significant.

WEIGHT AT 75 DAYS OF AGE: Average weight of femaleprogeny at 75 days of age was 175 grams for the early line and 179.8grams for the late line. Late-line females were thus 3.9 gramsheavier (Table 16). These values are the average of females raisedin first litters of the first, second, and third generations of each line.

Table 16. AVERAGE WEIGHT AT 75 DAYS OF AGE.

Early .........................................................................................Late :----------------------------------------------------------------------------------------

Difference ------------------------------------------------------------------------------

WeightGrams175.91798

lines .....................................................-line; .--...._..... ............................................

Line

Degreesof

freedom

1

1911

101

744.11192.o;

Numberfemales

11;0

200

F

3.86

0.04

Between Lines.Within lines ......................................................

1

388

Meansquare F

0.150.23

Total ------- ---------------------------------------------------- I 389

bah


The difference between lines lacked significance (Table 17).

Table 17. ANALYSIS OF VARIANCE OF WEIGHT AT 75 DAYS OF AGE.

MeanSource of variation square

BetweenWithin

Total

GROWTH RATE : The difference between female progeny ofthe two lines with respect to average gain per day from weaning totime of vaginal opening was 0.04 grams. Average rate of gain dur-ing this period was 3.87 grams per day for early-line females and3.83 grams per day for late-line females (Table 18).

Table 18. AVERAGE DAILY GAINS FROM WEANING TO TIME OFVAGINAL OPENING.

EarlyLate ----- -------------------------------------------------------------

Difference ----------------------------------------------------------------------------

Gain

Grams/day3.873.83

The average daily gains of the two lines did not differ sig-nificantly (Table 19).

Table 19. ANALYSIS OF VARIANCE OF GROWTH RATE.

Source of variation

Degreesof

freedom

0.63

Interrelation of factors influencing age at the time of vaginalopening.

The multiple correlation coefficient of age at the time of vaginalopening on five variates (number of young born, number of youngweaned, weaning weight, weight at the time of vaginal opening, and

PC.

TI

r

4. Meansquare

727.622.54

Gh l+"interval

!:L


gain per day from weaning to the time of vaginal opening) after lineeffects had been eliminated was significantly different from zero(R=0.89) by the analysis of variance (Table 20). The correspond-ing value obtained for R2 (0.79) indicates that 79 per cent of thevariability in age at the time of vaginal opening is accounted forby the variation in the five other variates studied.

Table 20. TEST OF SIGNIFICANCE OF THE MULTIPLE CORRELATIONCOEFFICIENT OF AGE AT TIME OF VAGINAL OPENING

ON THE FIVE OTHER VARIATES.

Source of variation

Regression ---------------------------------------------------------Residual --------------------------------------------------- ......--....---

Degreesof

freedom

Total ----------------------------------- --------- 388

* Significant.

Partial regression coefficients for age at the time of vaginalopening on each of the five variates were obtained while keepingthe remaining four variates constant. The five partial regressioncoefficients and their respective confidence intervals with a confidencecoefficient of 0.95 are presented in Table 21. For each additional ratborn in a litter (within the range of 2 to 15), the mean age at time ofvaginal opening of the females in the litter was increased by 0.08 days.For each additional rat weaned in a litter (within the range of 2 to 13),mean age at the time of vaginal opening of females in the litter was

Table 21. PARTIAL REGRESSION COEFFICIENTS AND THEIR RESPEC-TIVE CONFIDENCE INTERVALS OF AGE AT TIME OF VAGINAL

OPENING OF EACH OF THE FIVE OTHER VARIATES.

ariate

Partialregres-sion co-efficient onfidence

Days/numberof young

onfidence interval

Days/gram

Number born 00848 -0 0415 to 0.2146Number weaned -0.1012 -0.2180 to 0.0156Weaning weight -0.2305 -0.2586 to -0.2024Weight at time of

V.O . ........................... 02020 0.1896 to 0.2143Gain per day ................ -4.3046 -4.6673 to -3.9420

285.96*


decreased by 0.10 days. Each increase of 1 gram in weaning weight(within the range of 31 to 79 grams) reduced the mean age at thetime of vaginal opening 0.23 days. Each gram increase in weight attime of vaginal opening (within the range of 83 to 201 grams) in-creased the mean age at the time of vaginal opening 0.20 days. Ageat the time of vaginal opening was reduced 4.30 days for each gramper day increase in average rate of gain from weaning to time ofvaginal opening (within the range of 2.27 to 5.11 grams per day).

Partial regression coefficients were tested for significance by the"t" test. The t-values for age at time of vaginal opening on weaningweight, weight at time of vaginal opening and gain per day fromweaning to time of vaginal opening showed these three partial re-gression coefficients to be significantly different from zero (Table 22).

Table 22. TEST OF SIGNIFICANCE OF EACH OF THE FIVE PARTIALREGRESSION COEFFICIENTS

Variate

Degreesof

freedom

Number born --------- .............. 383 1.28Number weaned --------- _____----- 383 -1.70Weaning weight ............ ...... 383 -16.07*Weight at time of V O. ------- 383 31.96*Gain per day 383 -23 26*

* Significant.

Interrelation of factors influencing rate of gain of femalesfrom weaning to time of vaginal opening

The multiple correlation coefficient (R) of rate of gain on fiveother variates (number of young born, number of young weaned,weaning weight, weight at the time of vaginal opening, and age atthe time of vaginal opening) was 0.81 after line effects were elim-inated. This coefficient was found significantly different from zeroby the analysis of variance (Table 23). The corresponding valuefor R2 (0.66) indicates that 66 per cent of the variability in averagedaily gain from weaning to time of vaginal opening is accounted forby variation in the five other variates studied.

Partial regression coefficients for gain per day on each of thefive other variates were obtained while keeping the four remainingvariates constant. The five partial regression coefficients and theirrespective confidence intervals with a confidence coefficient of 0.95are presented in Table 24.

383

11.9455U.U8U-1

1;

ariale

--..- ...................... 38f3-

.f,


Table 23. TEST OF SIGNIFICANCE OF THE MULTIPLE CORRELATIONCOEFFICIENT OF RATE OF GAIN ON THE FIVE OTHER VARIATES.

Source of variation

Degreesof Mean

freedom square

Regression --------------------------------------- 148.54*Residual ------------------------ ---------------------------------------

Total

* Significant.

Table 24. PARTIAL REGRESSION COEFFICIENTS OF GAIN PER DAYON EACH OF THE FIVE OTHER VARIATES.

Partialregres-sion co-efficient onfidence interval

Grams per day/young

onfidence interval

Grams per day/dayNumber born ---------------- 0.0254 0.0024 to 0 0483Number weaned .--------- -0.0117 -0 0325 to 0.0091Weaning weight ---------- -0.0250 -0 0310 to -0 0190Weight at time of

V.0. ............................ 00318 0.0290 to 0.0346Age at time of V.0._-.. -0 1360 -0 1475 to -0 1246

Thus, for each additional rat born per litter (between therange of 2 to 15) the mean gain per day of females in the litterincreased 0.03 grams. For each additional rat weaned per litter(within the range of 2 to 13) the mean gain per day of females in thelitter decreased 0.01 grams. Each gram increase in weaning weight(within the range of 31 to 79 grams) decreased the mean gain perday 0.02 grams. Mean gain per day increased 0.03 grams for eachgram increase in weight at the time of vaginal opening (within therange of 83 to 201 grams). Each additional. day of age at time ofvaginal opening (within the range of 32 to 52 days) reduced meangain per day 0.14 grams.

Partial regression coefficients were tested for significance bythe "t" test. The t-value for number of young born per litter,weaning weight, weight at time of vaginal opening and age at timeof vaginal opening showed these four partial regression coefficientsto be significantly different from zero (Table 25).

I

............................................................................................................................

.............. ......................................... .........................................

................................... ----------------


Table 25. TEST OF SIGNIFICANCE OF EACH OF THE FIVE PARTIALREGRESSION COEFFICIENTS OF TABLE 24.

Variate

Degreesof

freedom

Number born ... 383 2.16*Number weaned 383 -110Weaning weight ............ 383 -8.22*Weight at time of V.O. 383 22.57*Age at time of V.0. .... 383 -23.26*

* Significant.

Discussion

The estimate of heritability of age at time of vaginal openingin this study was 17 per cent. This is the average of heritabilityestimates for each of three generations. The average estimate prob-ably is more reliable than any estimate for a single generation. Thisindicates that 17 per cent of variability in age at the time of vaginalopening is due to additive gene action. The remaining 83 per centis accounted for by environmental factors, nonadditive gene actionand environment-gene interactions. It is that portion (17 per cent)of the variability due to additive gene action that can be effectivelycontrolled by selection. While 17 per cent is not-considered a highheritability, it is enough to offer possibilities of improvement throughselection. That age at the time of vaginal opening is partially undergenetic control is substantiated by the fact that progeny of earlyand late lines were significantly different with regard to averageages at time of vaginal opening. This separated the two linesgenetically by selection. Stone and Barker (1940) also were ableto change age at time of pubescence in rats by selection. After sixgenerations of selection for early and late puberty in both femalesand males, females of two lines differed in age at time of pubertyby an average of 13.9 days. Blunn (1939) concluded that geneticcomposition was an important factor in determining age of pubertyin rats. The estimate of heritability of age at the time of vaginalopening in rats is in general agreement with estimates of heritabilityof characters associated with fertility in other species. Heritabilityof litter size in swine has been estimated to be 17 per cent (Stewart,1945), 22 per cent (Cummings, et al, 1947), and 24 per cent (Blunnand Barker, 1949). Heritability estimates for size of litter at wean-ing have been reported as 22 per cent (Blunn and Barker, 1949)and 32 per cent (Cummings, et al, 1947), while estimates for


heritability of litter weight at weaning have varied from 7 per cent(Cummings, et al, 1947) to 37 per cent (Blunn and Barker, 1949).

It should be noted that age at time of vaginal opening is onlyone indication of pubescence. This must be taken into accountespecially when making comparisons between species. Further, theestimate of heritability obtained in this study is subject to the usuallimitations in application and, like other heritability estimates, toerrors inherent in the method of estimation. In the case at hand,there are possibilities of over-estimate due to maternal effects andsex-linkage. On the other hand, a somewhat higher estimate wouldbe expected if the degree of heterozygosity in the rats studied wasmore and more comparable to that normally existing in farm animalssuch as swine, sheep, `and cattle.

It is of value to know what effect selection for a given charactermay have upon other characters. Since selection for early and lateage at the time of vaginal opening was successful in this study, itwas possible, by comparing performance of progeny of the two lines,to determine if selection for age at time of vaginal opening had bene-ficial, detrimental, or no effects on certain productive characters. Thecharacters studied in this connection were number of young bornper litter, number of young weaned per litter, weaning weight, weightat the time of vaginal opening, weight at 75 days of age, and averagedaily gains from weaning to time of vaginal opening. Comparisonof the two lines showed that early-line females gave birth to largerlitters. This greater fertility is probably the result of greater func-tional capacity of the ovary-more eggs shed. If this apparent asso-ciation between early puberty and high fertility is found in farmanimals, age of attaining puberty could be used as an additionalguide in improving fertility. Asdell (1946) is of the opinion thatsuch an association does exist in some species.

Number of young raised per litter and percentage of litterssuckled were the same for females of the two lines. Average wean-ing weight of late-line progeny was greater than that of early-lineprogeny. Since litters of early-line females were larger at birthit appears there is a negative relation of fertility to suckling ability.Asdell, et al (1941) found that weight of young at birth was in-versely related to size of litter and that birth weight advantage of youngin smaller litters was maintained through weaning time. Thus,heavier weaning weights of late-line progeny may result pri-marily from prenatal advantage which originates in the smallernumber of young rather than in any difference in total amount offood supplied by the females, either prenatally or postnatally. Thisdoes not explain why average numbe- of young weaned per litter


was the same in the two lines-even though average number bornwas greater in the early line. One explanation is that increasedgonadotropic hormone secretion by the anterior pituitary, coinci-dent with early puberty, results in the diminished output of lacto-genic hormone which reduces milk production. Asdell and Crowell(1935) have postulated a somewhat similar quantitative inverse re-lationship between the secretion of the gonadotropic and growthhormones. This, in itself, might explain the results obtained heresince no direct evidence of reduced milk secretion was obtained.

There was no difference in the two lines in rate of gain orweight at 75 days of age. Lack of difference in growth rate fromweaning to time of vaginal opening and weight at 75 days of ageindicates there is no difference between the two lines in potentialgrowth impetus, although Cole and Casady° (1947) reported that ratsin their early-pubescent strain were heavier at 85 to 90 days of agethan those in their late-pubescent strain.

Selection for several characters must normally be carried outsimultaneously. It is important, therefore, to know the effect thatcertain productive characters might have on the progress obtainedby selection for early pubescence. This information would be valu-able in showing genetic or physiological correlations of charactersand might help in predicting effectiveness of selection for two ormore characters at one time. The effects of number of young bornper litter, number of young weaned per litter, weaning weight, weightat time of vaginal opening, and average daily gain from weaning totime of vaginal opening on age at the time of vaginal opening weredetermined by use of multiple regression after the line effects wereeliminated. Effect of each of the five variates was determined bycomputing their respective partial regression coefficients while theremaining four variates were held constant statistically.

Neither the number of young born nor the number of youngweaned per litter affected age at time of vaginal opening. Thussize of litter in which a female was born or reared had no effecton her age at time of vaginal opening. This is in agreement withthe findings of Blunn (1939), but contrary to the reports of Engle,Crafts, and Zeithaml (1937), who found that females reared inlitters of 10 or 11 were significantly older at the time of vaginalopening than those reared in litters of 2 to 5.

The partial regression coefficients for weaning weight (-0.23days per gram) and daily gain (-4.30 days per gram) were sig-nificantly different from zero. Thus the heavier rats at weaningtime and those that gained the fastest after weaning were youngestat time of vaginal opening. Blunn (1939) also found a negative


correlation between daily gain from weaning time to the time ofvaginal opening and age at time of vaginal opening. Asdell andCrowell (1935) and Evans and Bishop (1922) have likewise foundevidence of a relationship between growth rate and rate of pubescence.

In the present study, each change of 1 gram in weaning weightaccompanied a change of 0.23 days in the mean age at the timeof vaginal opening. Each change of one gram in gain per day alsoresulted in a change of 4.30 days in mean age at time of vaginalopening. If one selects for high weaning weight and a maximumrate of gain simultaneously, therefore, he should thereby accomplisha reduction of the age at the time of vaginal opening-even thoughno deliberate selection for this trait is practiced.

The multiple correlation coefficient of age at the time of vaginalopening on the five variates studied was 0.89 (R2=0.79), indicatingthat 79 per cent of the sum of squares of the deviations from themean in age at the time of vaginal opening is accounted for byvariability in the five variates. Thus, if one selected for these fivecharacters simultaneously and applied the appropriate selection pres-sure to each, progress with respect to age at the time of vaginalopening would be expected.

Since average daily gain from weaning to time of vaginal open-ing was found to be so important in determining the mean age atthe time of vaginal opening (partial regression coefficient = -4.30),it seemed important to determine the extent to which the othercharacters studied were affecting the rate of gain. This was accom-plished by the use of multiple regressions. The multiple correlationcoefficient obtained was 0.81 and R2 was 0.66. This indicated thattwo-thirds of the variability in rate of gain was accounted for byvariability in the five other variates studied.

Fertility, as indicated by the number of young born per litter,was found to be positively associated with rate of gain after wean-ing; whereas suckling ability, as indicated by the number of youngraised per litter, was not. Thus, the larger the litter at birth thegreater were the gains after weaning (partial regression coefficient= 0.02 grams per day). This leads to the assumption that onlyafter weaning do individuals born in large litters overcome theprenatal disadvantage (Asdell, et al, 1941) of lighter birth weightcoincident with large litters. It may be seen from the study thatweaning weight is negatively correlated with subsequent daily gains.This strengthens the contention that young in larger litters have aprenatal disadvantage. The partial regression coefficient of dailygain on weaning weight was -0.02 grams per day. Thus for eachchange of 1 gram in weaning weight a change of 0.02 grams perday in the mean rate of gain would be expected.


ConclusionsHeritability of age of female rats at the time of vaginal

opening was estimated to be 17 per cent.Selection for the extremes in age at time of vaginal opening

was successful and genetically early and late lines were established.Two divergent lines also differed in average litter size at birth,

average weaning weight, and average weight at the time of vaginalopening.

An association between early puberty and prolificacy appar-ently exists. There is some indication that this association is be-tween early puberty and number of ova shed.

The multiple correlation coefficient for age at time of vaginalopening with five other variates studied was 0.89 (R2=0.79), indi-cating that much of the variability in age at time of vaginal openingis accounted for by variability in number of young born and raisedper litter, weaning weight, weight at the time of vaginal opening, anddaily gain from weaning to time of vaginal opening.

Partial regression coefficients show that age at time of vaginalopening is not affected by size of litter, either at birth or weaning,but is negatively correlated with weaning weight and daily gain.

Rate of gain from weaning to time of vaginal opening is in-fluenced by number of young born and raised in the litter, weaningweight and age and weight at time of vaginal opening.

Partial regression coefficients show that increases in numberof young born and weight at time of vaginal opening are associatedwith greater gains and that increases in weaning weight and age attime of vaginal opening are associated with decreased rate of gain.


Bibliography

Asdell, S A. Patterns of Mammalian Reproduction. Ithaca, ComstockPublishing Co. 1946 437p.

Asdell, S. A., Ralph Bogart and G. Sperling. The Influence of Age and Rateof Breeding Upon the Ability of the Female Rat to Reproduce andRaise Young. New York (Cornell) Agricultural Experiment StationMemoirs 238. 1941. 26p.

Asdell, S. A., and Mary F. Crowell. The Effect of Retarded Growth Upon theSexual Development of Rats Journal of Nutrition 10:13-24 1935.

Blandau, Richard J and William L. Money. The Attainment of Sexual Ma-turity in the Female Albino Rat as Determined by the CopulatoryResponse. Anatomical Record 86:197-215. 1943.

Blunn, Cecil T. The Age of Rats at Sexual Maturity as Determined byTheir Genetic Constitution. Anatomical Record 74:199-213. 1939.

Blunn, Cecil T., and Marvel L. Baker. Heritability Estimates of Sow Produc-tivity and Litter Performance. Journal of Animal Science 8:89-97.1949.

Bogart, R., G. Sperling, L. L. Barnes and S. A. Asdell. Influence of Repro-ductive Condition Upon Growth in the Female Rat. American Journalof Physiology 128 :355-371. 1940.

Cole, H. H. and R. B. Casady. Studies on Indices and Causes of Prolificacyin Rats. Endocrinology 41:119-126. 1947.

Cummings, J. N., L. M. Winters and H. A. Stewart. The Heritability ofSome Factors Affecting Productivity of Brood Sows. Journal ofAnimal Science 6 :297-304. 1947.

Engle, E. T., R. C. Crafts and C. E. Zeithaml. First Estrus in Rats in Rela-tion to Age, Weight and Length. Proceeding of the Society of Experi-mental Biology and Medicine 37 :427-432. 1937.

Evans, H. M. and K. S. Bishop. On the Relation Between Fertility and Nu-trition. The Ovulation Rhythm in the Rat on a Standard NutritionalRegime. Journal Metabolic Research 1 :319-334. 1922.

Evans, H. M. and K. S. Bishop. On the Relation Between Fertility and Nu-trition. The Ovulation Rhythm in the Rat on Inadequate NutritionalRegimes. Journal Metabolic Research 1:335. 1922.

Freudenberger, C. B. A Comparison of the Wistar Albino and the Long-Evans Hybrid Strain of the Norway Rat. American Journal Anatomy50:293-349. 1932.

Greenman, M. J. and F. L. Duhring. Breeding and Care of the Albino Ratfor Research Purposes. Wistar Institute 1-109. 1923.

Hetzer, H. O. and G. W. Brier. Extent to Which Type Differences AmongSwine Affect Litter Size. Proceedings American Society AnimalProduction 135-138. 1940.

King, Helen Dean., and H. H. Donaldson Life Processes and Size of theBody and Organs of the Gray Norway Rat During Ten Generationsin Captivity. American Anatomical Memoirs 14:1-106. 1929.

Krider, J. L., B. W. Fairbanks, W. E. Carrol, and E. Roberts. Effectivenessof Selecting for Rapid and Slow Growth Rate in Hampshire Swine.Journal of Animal Science 5:3-15. 1946.

Long, J. A. and A. M. Evans. On the Attainment of Sexual Maturity andthe Character of the First Estrous Cycle in the Rat. AnatomicalRecord 18:244. 1920.


MacArthur, J. W. Genetics of Body Size and Related Characters. II. Satel-lite Characters Associated with Body Size in Mice. AmericanNaturalist 78 :224-237. 1944.

McKenzie, Fred F. Growth and Reproduction in Swine. Missouri Agricul-tural Experiment Station Research Bulletin 118. 1928. p.67.

Phillips, Ralph W., and John H Zeller. Sexual Development in Small andLarge Types of Swine. Anatomical Record 85:387-400. 1943..

Slonaker, J. R. The Effect of Pubescence, Oestruation and Menopause of theVoluntary Activity in the Albino Rat. American Journal Physiology68:294-315. 1924.

Snedecor, G. W. Statistical Methods. 4th Ed. Ames, The Collegiate Press,Inc. 1946. 485p.

Stewart, H. A. Appraisal of Factors Affecting Prolificacy in Swine. Journalof Animal Science 4:359-366. 1945.

Stone, Calvin P., and R. G Barker. Change of the Age of Puberty in AlbinoRats by Selective Mating. Proceeding of the Society of ExperimentalBiology and Medicine 44:48-50. 1940.

Effect of Selection for Age at Time of Puberty On ...

Documents

Transcript of Effect of Selection for Age at Time of Puberty On ...